AI/ML
August 1, 2025
I drafted this blog post in 2023, but it’s surprisingly still relevant.
I’m allergic to buzzwords. I compulsively pop hype balloons even when it’s against my own economic best-interest.
I run a games studio that makes videogames for doctors, and I forbid use of the word ‘gamification’. I give talks to hype-craving audiences at SXSW and elsewhere about ‘cutting through metaverse hype with a scalpel’: Calling out these technologies as a fraud at worst and at best, a ill-advised assessment of where real technology improvements will take us.
And to be clear, although my company makes extensive use of AI to make better games, I don’t sell any AI products.
Nonetheless, I have to dispel a certain perception that I encounter frequently from folks that are justifiably burned-out of hype curves: “AI is just another hyped-up technology like NFTs and the Metaverse”.
Bottom-Line: It’s not, and there’s a simple test to prove it.
The Simple Test
A question I ask the NFT and the metaverse bros (posed as 3 questions but they’re all asking the same thing): What’s the killer app? What’s the one thing I need the metaverse for? What do NFTs let me do that I couldn’t do without NFTs?
Real answer: Nothing.
Now of course there’s a whole conversation leading up to that conclusion, with a bunch of lame, poorly-though-through examples, none of which hold up to any scrutiny. Each example is paper thin:
“…wait – I’m going to buy a pair of virtual Nike shoes for my cheesy virtual avatar – why?”
“NFTs let me own my in-game item forever? But when the game company shuts down the server, I can’t use it anyway – isn’t it worthless?”
“…Why do you even need NFTs for that?”
Eventually these conversations end in same the inevitable cop-out: “NFTs/metaverse is one of those technologies where you just have to build it, and then the applications will reveal themselves”. This is, of course, total nonsense. Technology doesn’t work that way. The emperor has no clothes.
GenAI is the opposite. After using ChatGPT 3 for a few hours in 2022, one quickly realized: “Wow. I could literally sit down for 4 days and do nothing but write down all of the ways that LLM technology can benefit society and I would only be scratching the surface’
Diagnosing diseases, tutoring, writing poetry, summarizing scientific studies, conducting meta-analyses, brainstorming, translation, customized explanation, writing unit tests, writing applications, solving difficult engineering problems… and these are just categories of value-creating solutions enabled by LLMs.
That’s one major difference (and a quick measure) between 1) a technology that is truly disruptive and massively value-creating (like GenAI and LLMs) and a 2) hype bubble.